NutriScan: An Estimation of Nutrition and Volume of Food Item through Monocular Image Analysis

With the growing concern over diet-related health challenges, there is an urgent need for innovative tools to simplify and enhance the way we monitor and understand food consumption. This study introduces an innovative image-based system to estimate food volume and nutritional content from single monocular images, leveraging advanced computer vision and deep learning techniques for accurate food classification, volume estimation, and nutritional profiling. The approach aims to offer a robust and adaptable solution for dietary assessment, personalized nutrition, and efficient health tracking. By integrating food classification using a pre-trained Inception V3 model with precise volume estimation through depth analysis and segmentation, the system achieves robust and dependable outcomes. A custom segmentation model isolates the food region, while calibrated depth information ensures reliable volume measurements. The proposed framework holds potential for applications in dietary monitoring and health management by providing an efficient, image-based solution for analyzing food consumption.